Risk and Its Part In Engineering Failure

Ever played the board game RISK? Even if you haven't, you've most likely heard of it. The game is aptly named because, regardless of the strategy implemented, winning the game involves some level of risk-taking (or else just some really stupid ignorant opponents).

This is no different in industry and the marketplace. Any new undertaking or project with unknown variables involves some level of risk. Unfortunately, the costs of failure in the real world are much more severe than being conquered and watching your friend nemesis to take over the world.

But in regards to safety and preventing costly failures and disasters, the question is this: At what point is a risk worth taking and when (if ever) is a loss considered acceptable?

Factors and Farmer Diagrams

Most systems require numbers to be analyzed and evaluated. The Farmer diagram and its associated calculations are the results of numbers applied to risk.

Farmer curves breakdown risk into two factors: probability and severity (cost). Probability is the likelihood an event will occur, and severity is the extremity of an event or a measure of its consequences.

Calculations typically involve multiplying the probability and severity factors together to come up with a resulting value (higher value = higher risk). They can also be compared on graphs called Farmer diagrams, plotted on an axis of severity v. probability (see left; Source: System Safety Engineering). Based on these values, new systems can be compared to other systems and are put into categories to define their associated risks as acceptable, not acceptable, and in some cases ALARP (as low as reasonably practicable).

The problem with linear risk calculation is that in most cases, severity plays a bigger role than probability. In other words, just like in RISK the game, straight numbers do not always make one scenario equivalent to another. For example, many extremely frequent small accidents are not necessarily as risky as one or two rare but catastrophic accidents. Most calculations involve some additional proportionality to balance this.

Decision Problems

When defining the limits of risk mitigation, the reality is that an infinite amount of time and money can be spent making a process or product safer. But the safety of personnel and of the environment must be balanced by economic feasibility. The problem then becomes finding this limit.

RISK games require a number of key decisions regarding limits and resources. In addition to underestimating the size of another player's army, putting all your own resources in one basket (territory) can be devastating.

Similarly, the allocation of safety-associated costs must be spread out throughout the life cycle of the process or product. In the case of a building or structure, this includes the resources and time spent during its design, during its construction, and during its operational life (maintenance). Failures and disasters can result from skimping out on any of these periods.

Ultimately, laws will never (and in my opinion should never try to) satisfy all the necessary specifications and requirements of individual projects. Understanding the necessary considerations has and should come from following good engineering principles, practices, and standards, which should be handed down to young engineers by experienced ones. Education is the key to a proper understanding and mitigation of risks and failures in a design or production process.

Quantitative vs. Qualitative

Risk management (as opposed to assessment) is more than just a numbers game. After all, a figure or statistic which may be rejected by an individual may be an "acceptable risk" in the eyes of a corporation. This is a deadly game to play, especially if some very negative consequences can be made economically "affordable" within a company's decision analysis.

Similarly, in a RISK game, a player must take into account the human factor. Even with a calculated strategy that quantitatively accounts for the unknowns associated with dice-rolling; a player can easily lose if his opponents play unpredictably and illogically.

This is why the human, societal, and political factors associated with a risk are just as important as the statistics. While a technician or statistician may be able to look at or analyze a bunch of numbers, it is the responsibility of the engineers and project managers to consider the implications of all the variables surrounding and involved in their project.

Sources:

ETH Swiss Federal Institute of Technology - Risk and Safety in Engineering (pdf)

ETH Swiss Federal Institute of Technology - Risk and Safety in Civil, Surveying and Environmental Engineering (pdf)

System Safety Engineering - Risk

University of Illinois - The Risk Assessment Methodology (pdf)